Generalized Deterministic-Random Tradeoff in Integrated Sensing and Communications: The Sensing-Optimal Operating Point

TL;DR

This paper investigates the fundamental tradeoff between sensing and communication in integrated systems, identifying the sensing-optimal point within the capacity-distortion region, especially for non-convex performance metrics.

Contribution

It extends understanding of the deterministic-random tradeoff by analyzing the sensing-optimal point for general and non-convex sensing metrics in ISAC systems.

Findings

DRT exists for non-convex sensing metrics

Sensing-optimal point characterized within capacity-distortion region

Analysis applies to detection probability in target detection

Abstract

Integrated sensing and communications (ISAC) has been recognized as a key component in the envisioned 6G communication systems. Understanding the fundamental performance tradeoff between sensing and communication functionalities is essential for designing practical cost-efficient ISAC systems. In this paper, we aim for augmenting the current understanding of the deterministic-random tradeoff (DRT) between sensing and communication, by analyzing the sensing-optimal operating point of the fundamental capacity-distortion region. We show that the DRT exists for generic sensing performance metrics that are in general not convex/concave in the ISAC waveform. Especially, we elaborate on a representative non-convex performance metric, namely the detection probability for target detection tasks.